[PATCH v2] ARM: kexec: selective MMU identity mapping

[linux@arm.linux.org.uk (Russell King - ARM Linux)]

On Wed, Feb 02, 2011 at 05:10:37PM -0500, Nicolas Pitre wrote:
> On Wed, 2 Feb 2011, Russell King - ARM Linux wrote:
> 
> > On Wed, Feb 02, 2011 at 03:43:29PM +0100, Per Fransson wrote:
> > > When restarting using the kernel kexec functionality the MMU
> > > needs to be turned off. Any code which does this needs to use
> > > identity mapped addresses to get reliable results. In the ARM
> > > kexec case this identity mapping is done:
> > > 
> > > - using the page table of the current task
> > > 
> > > - for all addresses normally used by user space,
> > >   i.e. 0x00000000-PAGE_OFFSET
> > > 
> > > If kexec is used at a kernel crash to collect a core dump this
> > > means that we lose important information.
> > > 
> > > This is what this patches does:
> > > 
> > > * Actually turns off the MMU, which has been omitted by mistake
> > > 
> > > * Sets up a more selective identity mapping
> > > 
> > > * Restores the old mapping once the MMU is off
> > > 
> > > Signed-off-by: Per Fransson <per.xx.fransson@stericsson.com>
> > > ---
> > > v2 changes:
> > > 
> > > * now uses (modified versions of) the identity mapping functions in idmap.c
> > >   as they look in 2.6.38-rc1. Some pud-level code has been added there in
> > >   linux-next.
> > 
> > Thanks.  As this been tested with ARMv4/v5 CPUs as well?
> 
> I have doubts about the effectiveness of setup_mm_for_reboot() on any 
> machine where physical RAM starts at 0xc0000000 or above, such as on 
> SA1100.

It's known to work on Assabet.  It works on SA1100 because the kernel
mapping is already a 1:1 mapping.

What setup_mm_for_reboot() is doing on Assabet though is making the
flash available for cpu_reset(0) to be able to call, not making the
kernel code for cpu_reset() available for calling.